EMBRYOGENESIS/SOMATIC EMBRYOGENESIS

Agrobacterium tumefaciens-mediated transformationof 'Brewster' ('Chen Tze') litchi (Litchi chinensis Sonn.)with the PISTILLATA cDNA in antisense

Guillermo Padilla & José A. Pérez & Irene Perea-Arango &

Pamela A. Moon & Miguel A. Gómez-Lim & Andrés A. Borges &

Marino Expósito-Rodríguez & Richard E. Litz

Received: 10 October 2012 /Accepted: 13 May 2013 / Editor: J. Forster# The Society for In Vitro Biology 2013

Abstract In order to address the need for production ofparthenocarpic fruits, embryogenic cultures of 'Brewster'('Chen Tze') litchi derived from leaves of a mature tree weretransformed with the PISTILLATA (PI) cDNA in antisenseorientation through Agrobacterium-mediated transformation.In four independently transformed lines, transgene integrationwas confirmed by conventional and quantitative PCR. Thefour transgenic lines differed with respect to the relative copynumber of the transgene and expression levels, and there wasno correlation between the two traits. The expression of thelitchi PI homolog was detected in transformed and non-transformed embryogenic cultures. Although expressionlevels were low overall, they were significantly higher inwild-type plants compared to lines that were transformed.This result suggests that post-transcriptional silencing of thelitchi PI homolog induced by an antisense oriented transgene

could be a successful strategy; however, silencing in floralprimordia to produce parthenocarpic fruits can only be con-firmed in mature plants after several years.

Keywords Litchi . Somatic embryogenesis .

Post-transcriptional silencing . Relative transgene copynumber . Gene expression

Introduction

Litchi (Litchi chinensis Sonn., family Sapindaceae) is asubtropical fruit tree native to South-East Asia. The fruit isa drupe, in which the edible part is the aril. Normal fruitscontain a single large seed that, along with the pericarp,reduces the flesh content within the fruit by 50%–70%

G. Padilla (*)Instituto de Productos Naturales y Agrobiología, Consejo Superiorde Investigaciones Científicas, Avda. Astrofísico FranciscoSánchez 3, 38206, La Laguna, Santa Cruz de Tenerife, Spaine-mail: gpadilla@cib.csic.es

A. A. Borges :M. Expósito-RodríguezConsejo Superior de Investigaciones Científicas, Instituto deProductos Naturales y Agrobiología, Avda. Astrofísico FranciscoSánchez 3, 38206, La Laguna, Santa Cruz de Tenerife, Spain

A. A. Borgese-mail: aborges@ipna.csic.es

M. Expósito-Rodrígueze-mail: mexpos@essex.ac.uk

J. A. PérezInstituto de Enfermedades Tropicales y Salud Pública de Canarias,Área de Genética, Universidad de La Laguna, Avda. AstrofísicoFrancisco Sánchez s/n, 38203, LaLaguna, SantaCruz deTenerife, Spaine-mail: joanpere@ull.edu.es

I. Perea-ArangoCentro de Investigación en Biotecnología, Universidad Autónomadel Estado de Morelos, Av. Universidad 1001, Col. Chamilpa,62209, Cuernavaca, Morelos, Mexicoe-mail: iperea@uaem.mx

P. A. Moon : R. E. LitzTropical Research and Education Center, University of Florida,18905 SW 280 Street, Homestead, Gainesville, FL 33031-3314, USA

P. A. Moone-mail: pamoon@ufl.edu

R. E. Litze-mail: relitz@ufl.edu

M. A. Gómez-LimUnidad Irapuato, CINVESTAV–IPN, Libramiento Norte CarreteraIrapuato-León, Km 9.6, AP 629, Irapuato 36500 Guanajuato, Mexicoe-mail: mgomez@ira.cinvestav.mx

In Vitro Cell.Dev.Biol.—PlantDOI 10.1007/s11627-013-9533-9

(Menzel et al. 2005). Two different types of seedless litchifruit occur: aborted-seeded fruit and parthenocarpic fruit.Aborted-seeded fruits, known as “chicken tongue”, arisefrom abnormal ovules or as a result of incomplete fertiliza-tion, whereas parthenocarpic seedless fruits develop whenthe ovary has not been fertilized (Huang 2005). The occur-rence of aborted-seeded fruits is cultivar dependent and canvary from season to season (Wu et al. 2007). Fruits withaborted seeds are preferred by consumers, therefore demand-ing a higher price in the market (Subhadrabandhu and Stern2005). There are a few cultivars that produce nearly seedlessfruit, although the fruits are smaller and usually weigh <10 g(Menzel 2002). Most 'Hexiachuan' fruits are parthenocarpic,although they have a small size and are of low quality (Wuet al. 2007). The expansion of litchi into other regions of theworld has been limited by its poor production due to irregularflowering and unpredictable weather that limits pollinationduring blooming (Huang et al. 2001; Litz and Raharjo 2005).It has been suggested that induction of seedless litchi fruitcould resolve this (Huang 2001).

Conventional breeding of litchi has several limitationsdue to its long juvenile period (7–8 yr), the length of timebefore horticultural traits can be evaluated, low fertility, andhigh levels of heterozygosity common to most tropical andsubtropical perennial fruit crops (Gomez-Lim and Litz2004). Most new cultivars are selected from superior seed-lings obtained after open pollination. Cross-breeding is notcommonly used because the genetic background of mostlitchi germplasm is unclear (Wu et al. 2007). Different strat-egies have been attempted in order to induce both aborted-seeded fruits and parthenocarpic fruits, by application ofgrowth regulators that inhibit zygotic embryo development(Huang and Stern 2005).

The application of biotechnology approaches for im-provement of existing litchi cultivars has been impeded bythe absence of a de novo regeneration pathway from maturephase or elite cultivars. Raharjo and Litz (2007) describedthe induction of embryogenic cultures from leaves of maturephase 'Brewster' ('Chen Tze') trees, a prerequisite for utiliz-ing genetic transformation and in vitro mutagenesis to im-prove this and other superior cultivars. In this study, we haveaddressed the genetic transformation of 'Brewster' ('ChenTze') that ultimately could result in parthenocarpy.

Floral regulatory genes were initially studied inAntirrhinum majus and Arabidopsis thaliana, and their func-tions were explained by the ABC model (Schwarz-Sommeret al. 1992; Weigel and Meyerowitz 1994). Although the twodicotyledonous species are distantly related, the set of regu-latory genes in both species is almost identical (Samach et al.1997). Homologous genes have been isolated from a widerange of plant species. In every case, the set of floral regula-tory genes belongs to the well-known MADS-box family,whose members contain a highly conserved DNA-binding

region, i.e., the MADS-box domain (Riechmann andMeyerowitz 1997). In A. thaliana, PISTILLATA (PI) is a Bclass gene that acts in whorls 2 and 3 during floral organdevelopment. Hence, the PI mutations convert petals tosepals and stamens to carpels (Bowman et al. 1991; Gotoand Meyerowitz 1994). Some parthenocarpic fruits, e.g.,apple (Malus domestica), have been shown to result fromnatural mutations in a homolog of the PI gene, demonstratingthe functional similarity between the homeotic genes of aperennial woody tree species and those of an annual plant forcontrolling floral organ development (Yao et al. 2001).

The antisense strategy has been one of the most referencedapproaches for gene silencing not only for the study of genefunction but for specific breeding purposes as well. Withinfruit crops, it has been applied to apple (Kotoda et al. 2006),pear (Gao et al. 2007), plum (Callahan and Scorza 2007),strawberry (Jiménez-Bermúdez et al. 2002), citrus (Wonget al. 2001), mango (Cruz-Hernández et al. 1997), and papaya(Magdalita et al. 2002). By transforming litchi with the A.thaliana PI cDNA in the antisense orientation, we expect tosuppress the homologous gene in litchi (PI-Lch) and ultimate-ly obtain parthenocarpic fruit development.

The goals of this study were to describe a procedure forgenetic transformation of an elite litchi cultivar 'Brewster'('Chen Tze') and to use a gene that could address the majorbreeding limitation of seedlessness, which would involveinterfering with stamen development and forcing fruit produc-tion without pollination. For these purposes, four differenttransgenic litchi lines that resulted from Agrobacterium-me-diated transformation of embryogenic cultures were studied.A relative quantification of the transgene was performed,along with the study of its expression and its ability to silencethe putative endogenous homologous gene in litchi.

Materials and methods

Plant material. An embryogenic culture of 'Brewster' ('ChenTze') L. chinensis Sonn. was induced from leaflets of youngvegetative flushes ofmature, field-grown trees from the tropicaland subtropical fruit collection of the University of FloridaTropical Research and Education Center, Florida, USA. The'Chen Tze' litchi is a traditional cultivar of South China that hasbeen vegetatively propagated for several centuries; it was in-troduced into Florida as air layered plants in the early twentiethcentury and renamed 'Brewster'. Therefore, the leaf explantswere from several hundred-year-old trees. The leaflets weresurface-sterilized with 0.02% (w/v) sodium hypochlorite for10 min and washed with three rinses of sterile, deionized water.Intact sterile leaflets were plated on induction medium in 125 ×15 mm sterile disposable Petri dishes. Induction mediumconsisted of B5 (Gamborg et al. 1968) major salts andMurashige and Skoog (MS;Murashige and Skoog 1962)minor

PADILLA ETAL.

salts and organic components with 100 mg L−1 myo-inositol,400 mg L−1 glutamine, 200 mg L−1 casein hydrolysate, 30 gL−1 sucrose, andwith 4.52mM2,4-dichlorophenoxyacetic acid(2,4-D), 9.30 μM kinetin, and 3 g L−1 Gel-GroTM (ICNBiochemicals, Santa Ana, CA) gellan gum (Raharjo and Litz2007). The kinetin was filter-sterilized and added to autoclavedmedium. The Petri dishes were sealed with Parafilm®(Pechiney Plastic Packaging Company, Chicago, IL).

Embryogenic cultures were maintained in liquid mediumcontaining half-strength MS major salts and full-strength MSminor salts and organic components, 30 g L−1 sucrose, 4.2μM2,4-D and 0.91 μM zeatin. The zeatin was filter-sterilized andadded to autoclaved maintenance medium. Embryogenic sus-pensions consisting of 200 mg of proembryogenic masses(PEMs) were inoculated into 40 mL of maintenance mediumin 125-mL Erlenmeyer flasks. The flasks were sealed withsterile, heavy-duty aluminum foil and Parafilm®. The cultureswere maintained at 100 rpm on a rotary shaker and weresubcultured at 2-week intervals into fresh medium. At eachsubculture, suspensions were passed through sterile, nylonfiltration fabric (1.6 mm opening), and an inoculum consistingof about 200 mg of the larger fraction size (>1.6 mm opening)was added to each 125-mL Erlenmeyer flask containing40 mL fresh liquid maintenance medium. Growth conditionsfor induction and maintenance were darkness at 25°C.

The pH of all media was adjusted to 5.7. Media weresterilized by autoclaving at 1.1 kg cm2 at 121°C for 20 min.

Bacterial strain and plasmid. The disarmed Agrobacteriumtumefaciens strain EHA105with the binary vector pCambia3301(http://www.cambia.org/daisy/cambia/2071/version/1/part/4/da-ta/pCAMBIA3301.pdf?branch=main&language=default;Cambia, Canberra, Australia) was used for genetic transforma-tion (Fig. 1). The transferred DNA region contains the barexpression cassette providing resistance to phosphinothricin(Block et al. 1987) driven by the Cauliflower Mosaic Virus(CaMV) 35S promoter, the PISTILLATA cDNA sequence(Goto and Meyerowitz 1994) in reverse orientation flanked bythe CaMV 35S promoter and the rbcS terminator sequences,and the GUS (uidA gene) expression cassette (Jefferson 1987)as a reporter gene. The vector was introduced into A.tumefaciens strain EHA105 by electroporation. The strainEHA105 containing pCAMBIA3301/B-PISTILLATA wasgrown in 1.0 mL LB broth at 27°C for 48 h. A single colonyof Agrobacterium strain EHA105/pCAMBIA3301/B-

PISTILLATA was inoculated into 5 mL of LB medium (pH7.0) supplemented with 50 mg L−1 kanamycin and 20 mg L−1

rifampicin at 28°C at 200 rpm in darkness for 16 h. The culturewas diluted to 20 mL with fresh LB medium modified by theaddition of 50 mg L−1 kanamycin, 20 mg L−1 rifampicin, and120 μM acetosyringone, and incubated at 125 rpm in darknessat 27°C. After 8 h, 0.8 mL of acetosyringone-activatedAgrobacterium suspension (OD600=0.2) was added to eachflask containing litchi embryogenic cultures.

Inhibition of PEM growth by phosphinothricin. The concen-tration of phosphinothricin that inhibited growth of litchiembryogenic suspension cultures was determined. The inoc-ulum consisted of embryogenic 'Brewster' cultures from fiveflasks (40 mL liquid maintenance media in 125-mLErlenmeyer flasks). Six concentrations of phosphinothricinwere tested in addition to the control: 0.5, 1.0, 1.5, 2.0, 2.5,and 3.0 mg L−1 with five replications for each treatment andthe control. Filter-sterilized phosphinothricin was added to40 mL liquid maintenance medium in 125-mL Erlenmeyerflasks, and 1.0 mL of settled cell volume (SCV) of PEMswas inoculated into each flask. The flasks were capped withaluminum foil and sealed with Parafilm®. Cultures weremaintained on a rotary shaker at 125 rpm at 27°C underdiffuse light. The SCV of PEMs was measured after 5 wkusing the method described above. For SCV measurement,PEMs were decanted into sterile graduated plastic centrifugetubes and the SCV was measured after about 1 min. Datawere analyzed by ANOVA at a 1% confidence level (SASInstitute, Cary, NC).

Co-cultivation with Agrobacterium. Approximately 4.0 g ofembryogenic 'Brewster' ('Chen Tze') cultures were sievedthrough sterile nylon filtration fabric (mesh size 1.8 mm), afterwhich 0.25 g of the filtrate was added to 125-mL flaskscontaining 40 mL of maintenance medium and 15 mg carbo-rundum. After 1 wk, the flasks were homogenized using avortex 30 s and 0.8 mL of log phase acetosyringone-activatedEHA105/pCAMBIA3301/B-PISTILLATA culture was addedto each flask. The PEMs and EHA105/pCAMBIA3301/B-PISTILLATA cultures were co-cultivated at 27°C for 3 d at100 rpm in darkness.

The PEMs were then transferred to 40 mL liquid mainte-nance medium supplemented with filter-sterilized 200 mg L−1

cefotaxime and 500 mg L−1 carbenicillin in 125-mL

Figure 1. pCAMBIA3301/B-PISTILLATA. The transferred DNA re-gion contains the bar expression cassette providing phosphinothricinresistance and driven by the CaMV 35S promoter, the PISTILLATA

sequence flanked by the CaMV 35S promoter and the rbcS terminatorsequences, and the GUS (uidA) expression cassette as a reporter gene.

AGROBACTERIUM TUMEFACIENS

Erlenmeyer flasks at 27°C and 125–130 rpm to eliminate A.tumefaciens. After 1 wk, the PEMs were transferred into freshliquid maintenance medium supplemented with filter-sterilized 200 mg L−1 cefotaxime, 500 mg L−1 carbenicillin,and 3.0 mg L−1 phosphinothricin, and subcultured at 2-wkintervals in liquid medium of the same formulation for 3 mo.

Growth of putative transgenic embryogenic cultures on se-lection medium. The growth of putatively transformed em-bryogenic cultures and non-transformed cultures on semi-solid selection maintenance medium with 2.0 mg L−1

phosphinothricin was compared. Each sample consisted of150±5 mg PEMs. There were four replicates for each treat-ment and the experiment was repeated twice. Weight mea-surements and subcultures were performed at 2-wk intervalsfor 8 wk.

Development of transformed somatic embryos. Somatic em-bryo development was initiated from putatively transformed em-bryogenic cultures from liquid maintenance selection mediumaccording to Raharjo and Litz (2007). Each suspension culturewas passed through sterile, nylon filtration fabric 2 wk aftersubculture, and the filtrate was transferred to fresh medium.Approximately 200 mg of PEMs were spread evenly on thesurface of semi-solid MS medium in 100 × 20 mm Petri dishescontaining 45 g L−1 sucrose, 10% (v/v) filter-sterilized coconutwater (CW), and 3 g L−1 gellan gum. CW was obtained fromfreshly harvested immature coconuts and was filter-sterilizedbefore adding to autoclaved medium. The cultures were incubat-ed in darkness at 25°C. Germination/conversion occurred follow-ing the transfer of mature, white-opaque somatic embryos ≥5mmin diameter to plant recovery medium, consisting of semi-solidMS medium with 30 mg L−1 sucrose, 14.4 μM gibberellic acid(GA3), and 3 g L−1 gellan gum in 100 × 20 mm Petri dishescontaining 50 mL of medium (Raharjo and Litz 2007). The GA3

was filter-sterilized and added to autoclaved medium. Followingroot and shoot emergence, plantlets were transferred individuallyinto sterile glass containers (90 × 55mm) containing 60mL semi-solid MS basal medium with 30 g L−1 sucrose, 100 mg L−1

activated charcoal, and 4 g L−1 Agargel (Sigma, St. Louis,MO). Growing conditions were 25°C with a 16-h photoperiod(60–80 μmol s−1 m−2) provided by cool white fluorescent lights.

GUS assay. Phosphinothricin-resistant PEMs, somatic em-bryos, and leaves of regenerated plants were evaluated forGUS (β-glucuronidase) activity. Tissue samples were washedwith sterile deionized water and treated with X-Gluc (5-bro-mo-4-chloro-3-indolyl B-D glucuronide; Sigma) accordingto Jefferson (1987). The tissues were cleared withmethanol/acetone (3:1) and stored in 10% (v/v) glycerol.

DNA and RNA analyses. In longan (Dimocarpus longanLour.), a closely related member of the Sapindaceae family,

a PISTILLATA-like gene (LVFS-199) expressed in floralbuds has been described (Matsumoto 2006). The codingsequence of LVFS-199 (GenBank Acc DT368224) has a96% identity with the one defined as “Litchi chinensisMAD box protein 1 (MAD1) mRNA” in GenBank(AY705793). The litchi sequence is referred as PI-Lch.

Three independent purifications of genomic DNA andRNA were extracted from each putatively transformed lineand non-transformed PEMs. Nucleic acids were isolatedfrom 50 mg tissue samples using the E.Z.N.A.® SP PlantDNA Kit and the E.Z.N.A.® Plant RNA Kit (Omega Bio-Tek, Norcross, GA). The concentration of DNA samples wasdetermined with a spectrophotometer and then equalized to5 ng/μL with 0.5× TE buffer. RNA samples were furtherpurified by DNase I digestion, acid phenol/chloroform ex-traction, and ethanol precipitation (Nolan et al. 2006). RNAintegrity was assessed using the Agilent 2100 Bioanalyzer(Agilent Technologies, Palo Alto, CA). Only RNA sampleswith a RNA integrity number (RIN) above 5 were used forexpression analyses. cDNA was synthesized from 5 μg ofRNAwith the Transcriptor First Strand cDNA Synthesis Kit(Roche, Penzberg, Germany) using anchored oligo(dT)18primers. cDNA samples were diluted to a final volume of400 μL using 0.5× TE buffer.

Standard PCR was used to verify the presence of therecombinant DNA construct in PEMs. Three primer pairswere employed for the detection of the bar gene, uidA gene,and the antisense A. thaliana PISTILLATA sequence (Table 1).Each amplification reaction was carried out in a final volumeof 25 μL, containing 100 ng of gDNA and using theReadyMixTM Taq PCR Reaction Mix with MgCl2 (Sigma-Aldrich) according to manufacturer ’s instructions.pCambia3301 DNA was included as a positive control.PCRs were incubated in the GeneAmp® PCR System 2700(Applied Biosystems, Foster City, CA) with the followingtemperature profile: an initial step of 3 min at 94°C followedby 35 cycles of 1 min at 94°C, 1 min at 60°C, and 1 min at72°C, then a final step of 5 min at 72°C. Amplicons wereseparated on a 0.8% agarose gel in TBE buffer and stainedwith ethidium bromide.

Quantitative PCR (qPCR) was used for transgene copynumber estimation and gene expression analysis in PEMs.Amplification reactions were run in triplicate on theLightCycler® 480 platform (Roche), in a total volume of20 μL containing 6 μL of gDNA or cDNA template, 2 μMofeach amplification primer (Table 1), and 10 μL of 2×LightCycler 480 SYBR Green I Master (Roche). Amplifiedsequences for both PI and PI-Lch (156 bp and 171 bp,respectively) were included within the C-terminal domain,which shows the highest variability among the MADS-boxfamily members. The cycling conditions were as follows: aninitial denaturation step of 95°C for 10 min to activate theTaq DNA polymerase, followed by 45 cycles of denaturation

PADILLA ETAL.

at 95°C for 15 s, annealing at 60°C for 30 s, and extension at72°C for 30 s. The amplification process was followed by amelting curve analysis, ranging from 60 to 95°C, with a ramprate of 0.02°C/s and 25 fluorescence acquisitions per degreeCelsius. The quantification cycle (Cq) values were deter-mined with the LightCycler 480 Software 1.5, previousmanual setting of the fluorescence threshold. In order tocontrol a possible gDNA contamination of cDNA samples,each reverse transcription-quantitative PCR (RT-qPCR) as-say was accompanied by a RT-minus amplification reactionand only differences ≥10 cycles between the respective Cqvalues were accepted for further analyses. Besides calcula-tion of the melting temperature of each amplicon, amplifica-tion specificity was confirmed by agarose electrophoresis ina random subset of amplification reactions and sequencing ofsome selected amplicons (BigDye® Terminator v3.1 CycleSequencing Kit and 3500 Series Genetic Analyzer; AppliedBiosystems).

Relative quantities of the transgene or the targetedmRNAs, and the corresponding significance levels (p value),were calculated from Cq data following an efficiency-correction model (Pfaffl 2001) as implemented in the 2008version of REST software (Pfaffl et al. 2002). The PCRefficiency of each amplification reaction was estimated withthe DART-PCR workbook (Peirson et al. 2003). The SSRlocus named LMLY11, a single copy sequence of the litchi

genome (Viruel and Hormaza 2004), was used as referencegene to control gDNA input during transgene copy numberestimation. Normalization factors for mRNA level quantifi-cation were calculated as the geometric mean of the expres-sion levels of several control genes, whose suitability asinternal references was assessed with the geNorm software(Vandesompele et al. 2002). These reference genes wereselected among those proposed by Zhong et al. (2011).

Results

Verification of transformation. Embryogenic 'Brewster'('Chen Tze') litchi cultures were transformed with a DNAconstruct containing the bar gene, the uidA gene, and the A.thaliana PI cDNA in antisense. The growth of embryogeniclitchi suspension cultures was completely suppressed by1.0–2.0 mg L−1 phosphinothricin (Fig. 2), and the effectwas highly significant (p<0.01). Therefore, 2.0 mg L−1

phosphinothricin was used to select for transformed embryo-genic cultures that contained the bar gene.

Three months after co-culturing 'Brewster' ('Chen Tze')embryogenic cultures with EHA105/pCAMBIA3301/B-PISTILLATA and with continuous selection in the presenceof 2.0 mg L−1 phosphinothricin, transformation of PEMs wasverified on the basis of the X-Gluc histochemical reaction for

Table 1. PCR primers used for conventional and quantitative PCR

Gene name Gene description Primer sequence (5′–3′) Size(bp)

Reference

Conventional PCR

PI (415) Arabidopsis thaliana PISTILLATA F CTTCCATGGATCTTGGTGCT 415 (*)

R AGATTTGGCTGAATCGGTTG

uidA Beta-glucuronidase F CAGCGAAGAGGCAGTCAACGGGGAA 683 Raharjo et al. 2008

R CATTGTTTGCCTCCCTGCTGCGGTT

bar Resistance to phosphinothricin F GTCTGCACCATGGTCAACC 450 Ribas et al. 2006

R GAAGTCCAGCTGCCAGAAAC

Quantitative PCR

PI (156) Arabidopsis thaliana PISTILLATA F GAGGAGAAATGAGAAGATGATGG 156 (*)

R AGATTTGGCTGAATCGGTTG

LMLY11 Single copy gene in Litchi chinensis genome F CCCCACGAAAACTGAAGAGAG 206 (*)

R CATCAACGGTCCATCCTCC

PI-Lch Litchi chinensis MAD box protein 1 (MAD1)mRNA

F GGATGAGGGAAAATCTAAAGGAG 171 (*)

R GCATCTGAGGGTTGTAGTCATTG

GAPDH Glyceraldehyde-3-phosphate-dehydrogenase F GATACAGTTCCCGTGTTGTTGAC 121 Zhong et al. 2011

R CATAAAGACACATAACACCACACTC

EF-1α Elongation factor 1-alpha F CGTGGATTCGTCGCATCTAAC 183 Zhong et al. 2011

R CGTCCTATCATCAACAAAAGTGAC

RPII RNA polymerase-II transcription factor F TAAAGGATGATTTCTTGGGTGTG 168 Zhong et al. 2011

R CGTCCTATCATCAACAAAAGTGAC

(*) Primers designed in this study

AGROBACTERIUM TUMEFACIENS

β-glucuronidase activity. All transformed lines (T1–T4)showed intense GUS staining (Fig. 3a). Somatic embryosthat developed from T1 and T2 cultures were also positivefor GUS expression (Fig. 3b–d). At the time of writing, T2embryos had germinated (Fig. 4a–d).

The growth of the transformed embryogenic cultures onsemi-solid medium with 2.0 mg L−1 phosphinothricin (PPT)

was determined after 2, 4, 6, and 8 wk of incubation (Fig. 5).Throughout the measuring period, this growth was lowerthan the growth of non-transformed cultures in standardmaintenance medium. This was probably due to the strongselection for bar gene expression imposed on transformedcells since this gene encodes for PPT-acetyltransferase activ-ity necessary for PPT inactivation. Nevertheless, non-transformed cultures showed no significant growth on selec-tion medium with PPT (Fig. 5).

The presence of bar and uidA transgenes, and the PI anti-sense cDNA in transformed lines was confirmed by standardPCR. As shown in Fig. 6, amplicons with the expected size(Table 1) derived from bar, uidA, and PI sequences wereobtainedwith gDNA templates of the four selected transformedPEMs, whereas PEMs of non-transformed cultures did notshow any amplification product.

Transgene relative copy number. To determine the relativecopy number of the construct in the four transgenic lines,qPCR was performed with three independent gDNA samplesfrom each transformed line. Transgenic line 2 (T2) showedthe lowest amplification signal with PI (156) primers(Table 1) and it was used as calibrator, with an assigned

Figure 2. Effect of phosphinothricin on inhibition of growth of 'Brewster'('Chen Tze') litchi embryogenic suspension cultures.

Figure 3. X-Gluchistochemical reaction inembryogenic cultures (a) andsomatic embryos (b–d) of'Brewster' ('Chen Tze') litchi.

PADILLA ETAL.

value of 1. Gene copy number in T1 was 2-fold higher thanin T2, while in T3 and T4 the increment was 1.6-fold relativeto the T2 line (Fig. 7). All three differences with respect to T2were highly significant.

Expression analysis of transgene and endogenousPISTILLATA homolog. Initially, the suitability of GAPDH,EF-1α, and RPII housekeeping genes were tested as internalreferences for normalization of mRNA level measures.Analysis of pairwise variation between two sequential nor-malization factors with the geNorm software revealed thatthe EF-1α and RPII genes were sufficient for accuratemRNA quantification. The mean M and CV values for thepair EF-1α/RPII in our set of cDNA samples wereM=0.416

and CV=0.144. These values of expression stability areinside the ranges M≤0.5 and CV≤0.25, which have beenproposed by Hellemans et al. (2007) as acceptable for rela-tively homogeneous sample panels.

Transgene expression in all the transformed lines wasmeasured. The specificity of this analysis was confirmedby amplicon sequencing. The A. thaliana PI cDNA expres-sion level was significantly lower in line T1 than in the othertransgenic lines, whose expression values were estimated asfold changes over T1. Line T3 showed an expression level

Figure 4. The GUS (uidA) response in different transgenic leaves (a–d) from T2 line plantlet (e).

Figure 5. Growth of transformed proembryogenic cultures compared tonon-transformed cultures on semi-solid selection mediumwith 2.0 mg L−1

phosphinothricin. Results shown are the average of transformed plantsT1–T4.

Figure 6. Detection of the bar (a), uidA (b), and PI (c) genes by PCRanalysis of transformed PEMs (lines 1–4; transformed plants T1–T4, respec-tively).Mmolecular weight ladder (Promega, Ladder 100 bp, ref. PRG2101),C− control without DNA template, U non-transformed PEM (negativecontrol), P plasmid pCambia3301/B-PISTILLATA (positive control).

AGROBACTERIUM TUMEFACIENS

ca. 9-fold higher than T1. For lines T2 and T4, the averageincreases were 5-fold and 3-fold, respectively (Fig. 8).

PI-Lch mRNA in PEMs was detected as revealed byamplicon sequencing; however, the expression levels of thePI-Lch gene in these tissues were very low (Cq values around35–38) and consequently expressionmeasures suffered from alack of precision. In order to increase accuracy, amplificationreactions were performed on six replicates during expression

analysis of this endogenous gene. Despite this technical im-provement, comparisons involving the T1 transformed linewere precluded by the fact that only one out of three indepen-dent cDNA preparations from the same T1 PEM gave anamplification signal. The non-transformed line (WT)was usedas calibrator for the analysis of PI-Lch expression levelsshowed in Fig. 8. Compared to WT, PI-Lch mRNA levels inT2 and T4 decreased 7-fold, whereas in T3, the reduction wasca. 1.5-fold. For line T1, PI-Lch expression level was so lowthat it could not be determined with a minimal accuracy.Expression levels were significantly lower in the transgeniclines T2 and T4 compared to WT, and highly significant forT4. These results suggest that, to a variable extent, post-transcriptional silencing of the PI-Lch gene by the A. thalianaPI antisense mRNAwas successful. Interestingly, the amountof antisense mRNA did not correlate with the level of PI-Lchsilencing (Fig. 8).

Discussion

The most important tropical fruit crops, including banana(Ganapathi et al. 2001), papaya (Cheng et al. 1996; Yinget al. 1999), pineapple (Firoozabady and Gutterson 1998;Firoozabady et al. 2006), avocado (Cruz-Hernández et al.1998), and mango (Mathews et al. 1992, 1993), have beentransformed using Agrobacterium-mediated protocols basedon embryogenic cultures and the recovery of somatic em-bryos. Das and Rahman (2012) reported Agrobacterium-mediated transformation of litchi zygotic embryos with theregeneration of transgenic plantlets. In the present study,transgenic somatic embryos and plants from elite 'Brewster'('Chen Tze') litchi have been obtained.

A. tumefaciens strain EHA105 has been successfully usedfor the transformation of many different fruit tree crops, whiletransformation with pCambia as the binary plasmid vector hasonly been reported for kiwifruit (Actinidia deliciosa 'Hayward';Kim et al. 2008). Under our conditions, transformation efficien-cy based upon recovery of plants was low (<5%), althoughrecovery of transformed PEMs and somatic embryos was muchhigher (>80%). Raharjo and Litz (2007) also observed a veryhigh rate of germination/conversion for (non-transformed)'Brewster' ('Chen Tze') litchi somatic embryos. Similar resultswere observed with respect to plant recovery following transfor-mation of avocado embryogenic cultures (Raharjo et al. 2008).Somatic embryo development has occurred with two transgeniclines, T1 and T2, and plant recovery only with line T2.

Agrobacterium-mediated transformation usually results ina few copies of the transgenes integrated at a single locus as atandem array (Jorgensen et al. 1987). Most frequently, trans-gene copy number in transformed plants has been estimated bySouthern blot analysis. Determination of transgene copy

Figure 7. Relative copy number of A. thaliana PI cDNA in four transgeniclines (T1–T4). The LMLY11 gene was used for calculating normalizationfactors. T2 was used as calibration sample. Error bars indicate confidenceintervals (95%) estimated from a randomization assay with n=3. **Datahighly significantly different (p<0.01) from calibrator.

Figure 8. Relative expression level of A. thaliana PI antisense mRNA(PI) in four transgenic lines (T1–T4) and relative expression level ofLitchi PISTILLATA (PI-Lch) in the same transgenic lines and the wildtype (WT). EF-1α and RPII genes were used for calculating normali-zation factors. For PI analyses, line T1 was used as the calibrationsample, whereas for PI-Lch the calibrator was the WT line. Error barsindicate confidence intervals (95%) estimated from a randomization testwith n=3. *Data significantly different (p<0.05) and **highly signifi-cantly different (p<0.01) from the respective calibrator.

PADILLA ETAL.

number by real-time qPCR has become an effective alternativeto Southern blot (Shepherd et al. 2009), and in recent yearsseveral studies have demonstrated its effectiveness in compar-ison with Southern blot hybridization, particularly with woodyperennial species (Omar et al. 2008; Casu et al. 2012; Gadaletaet al. 2012). Using qRT-PCR in this study, it has been inferredthat transgene copies in T1was double that of T2, while T3 andT4 had about 1.6 more insertions than T2 (Fig. 7). These resultssuggest that the most probable absolute number of insertions istwo for T2, three for T3 and T4, and four for T1.

The expression of the PI antisense cDNA did not correlatewith the estimated number of copies in the obtained transformedlines (Figs. 7 and 8), as has been previously reported (Hobbset al. 1993; Gelvin 2003). For example, T1 showed the highestrelative number of copies among transformed lines even thoughit had the lowest PI expression level. In the same way, the levelof PI antisense mRNA in T3 was roughly 3-fold greater than inT4 (p<0.05), although their transgene copy numbers were notsignificantly different. Variable expression of plant transgeneshas been explained by position and epigenetic effects, as well asby post-transcriptional silencing (Matzke and Matzke 1998;Meins 2000). Post-transcriptional silencing is frequently associ-ated with multiple transgene copies within a cell, especiallywhen they are inserted in the genome as inverted repeats(Stam et al. 1997; Gelvin 2003).

PISTILLATAwas originally identified as a floral homeoticgene in A. thaliana (Bowman et al. 1989; Meyerowitz et al.1989). Although its highest expression occurs during earlyfloral development (Goto and Meyerowitz 1994), PI is alsoexpressed at low levels in young seedlings, embryos, and evenembryonic culture tissues (Kinoshita et al. 2001; Lehti-Shiuet al. 2005). This suggests that the gene may play an importantregulatory function during embryo development, in additionto a role in floral identity (Lehti-Shiu et al. 2005). In this study,we have also detected the expression of the putative litchi PIhomolog at very low levels. Although we demonstrated thatPI antisense is able to significantly decrease levels of PI-LchmRNA, we cannot conclude that silencing will be effectiveduring flower development, when the expression level ofendogenous genes is notably up-regulated in comparison toembryonic stages. Mature plants derived from T1 and T2somatic embryos will be available in 7–8 yr (typical juvenileperiod for litchi) and will allow us to confirm whether PI-LchmRNA levels will be sufficiently decreased in floral primor-dial to generate parthenocarpic fruits.

Since PI and PI-Lch cDNAs show a moderate identity(65%) biased towards the 5′-half of the sequence where theconserved MADS domain is encoded, we cannot totally ex-clude the possibility of the PI antisense sequence affecting theexpression of other members of MADS-box gene family;however, the fact that we were able to obtain plants fromgerminated transgenic somatic embryos suggests that any es-sential function for early development was not compromised.

Acknowledgment G. Padilla acknowledges financial support from theCabildo Insular de La Palma.

References

Block MD, Botterman J, Vandewiele M, Dockx J, Thoen C, Gossele V,Movva NR, Thompson C, Vanmontagu M, Leemans J (1987)Engineering herbicide resistance in plants by expression of adetoxifying enzyme. EMBO J 6:2513–2518

Bowman JL, Smyth DR, Meyerowitz EM (1989) Genes directingflower development in Arabidopsis. Plant Cell 1:37–52

Bowman JL, Smyth DR, Meyerowitz EM (1991) Genetic interactionsamong floral homeotic genes of Arabidopsis. Development 112:1–20

Callahan A, Scorza R (2007) Effects of a peach antisense ACC oxidasegene on plum fruit quality. Acta Horticult 738:567–573

Casu RE, Selivanova A, Perroux JM (2012) High-throughput assess-ment of transgene copy number in sugarcane using real-timequantitative PCR. Plant Cell Rep 31:167–177

Cheng YH, Yang JS, Yeh SD (1996) Efficient transformation of papayaby coat protein gene of papaya ringspot virus mediated byAgrobacterium following liquid-phase wounding of embryogenictissues with carborundum. Plant Cell Rep 16:127–132

Cruz-Hernández A, Gómez-Lim M, Litz RE (1997) Transformation ofmango somatic embryos. Acta Horticult 455:292–298

Cruz-Hernández A, Litz RE, Witjaksono, Gómez-Lim M (1998)Agrobacterium tumefaciens-mediated transformation of embryo-genic avocado cultures and regeneration of somatic embryos. PlantCell Rep 17:497–503

Das DK, Rahman A (2012) Expression of a rice chitinase gene en-hances antifungal response in transgenic litchi (cv. Bedana). PlantCell Tissue Organ Cult 109:315–325

Firoozabady E, Gutterson N (1998) Genetically transformed pineappleplants and methods for their production. International PatentApplication No. PCT/US98/03681

Firoozabady E, Heckert M, Gutterson N (2006) Transformation andregeneration of pineapple. Plant Cell Tissue Organ Cult 84:1–16

Gadaleta A, Giancaspro A, Cardone MF, Blanco A (2012) Real-timePCR for the detection of precise transgene copy number in durumwheat. Cell Mol Biol Lett 16:652–668

Gamborg OL, Miller RA, Ojima K (1968) Plant cell cultures. I. Nutrientrequirements of suspension cultures of soybean root cells. ExpCell Res 50:151–158

Ganapathi TR, Higgs NS, Balint-Kurti PJ, Arntzen CJ, May GD, VanEck JM (2001) Agrobacterium-mediated transformation of em-bryogenic cell suspensions of the banana cultivar Rasthali(AAB). Plant Cell Rep 20:157–162

Gao M, Matsuta N, Murayama H, Toyomasu T, Mitsuhashi W, DandekarAM, Tao R, Nishimura K (2007) Gene expression and ethyleneproduction in transgenic pear (Pyruscommunis cv. 'LaFrance') withsense or antisense cDNA encoding ACC oxidase. Plant Sci 173:32–42

Gelvin SB (2003) Agrobacterium-mediated plant transformation: the biologybehind the “gene-jockeying” tool. Microbiol Mol Biol Rev 67:16–37

Gomez-Lim MA, Litz RE (2004) Genetic transformation of perennialtropical fruit crops. In Vitro Cell Dev Biol Plant 40:442–449

Goto K,Meyerowitz EM (1994) Function and regulation of the Arabidopsisfloral homeotic gene PISTILLATA. Genes Dev 8:1548–1560

Hellemans J, Mortier G, De Paepe A, Speleman F, Vandesompele J(2007) qBase relative quantification framework and software formanagement and automated analysis of real-time quantitative PCRdata. Genome Biol 8:R19

Hobbs SLA, Warkentin TD, DeLong CMO (1993) Transgene copynumber can be positively or negatively associated with transgeneexpression. Plant Mol Biol 21:17–26

AGROBACTERIUM TUMEFACIENS

Huang H, Stern AS (2005) Fruit set, development and maturation. In:Menzel CM, Waite GK (eds) Litchi and longan: botany, produc-tion and uses. CABI, Wallingford, pp 115–140

Huang HB (2001) Towards a better insight into the development of thearilate fruit of litchi and longan. Acta Horticult 558:185–192

Huang X (2005) Fruit disorders. In: Menzel CM,Waite GK (eds) Litchi andlongan: botany, production and uses. CABI, Wallingford, pp 141–151

Huang XM, Li JG, Wang HC, Huang HB, Gao FF (2001) The relation-ship between fruit cracking and calcium in litchi pericarp. ActaHorticult 558:209–216

Jefferson RA (1987) Assaying chimeric genes in plants: the GUS fusionsystem. Plant Mol Biol Report 5:387–405

Jiménez-Bermúdez S, Redondo-Nevado J, Munoz-Blanco J, Caballero JL,Lopez-Aranda JM, Valpuesta V, Pliego-Alfaro F, Quesada MA,Mercado JA (2002) Manipulation of strawberry fruit softening byantisense expression of a pectate lyase gene. Plant Physiol 128:751–759

Jorgensen R, Snyder C, Jones JDG (1987) T-DNA is organized pre-dominantly in inverted repeat structures in plants transformed withAgrobacterium tumefaciens C58 derivatives. Mol Gen Genet207:471–477

Kim M, Kim SC, Jang KC, Song EY, Kang SJ, Song KJ (2008)Agrobacterium-mediated transformation of carotenoid synthesis genein kiwifruit (Actinidia deliciosa 'Hayward'). Acta Horticult 773:41–145

Kinoshita T, Harada JJ, Goldberg RB, Fischer RL (2001) Polycombrepression of flowering during early plant development. Proc NatlAcad Sci U S A 98:14156–14161

Kotoda N, Iwanami H, Takahashi S, Abe K (2006) Antisense expres-sion of MdTFL1, a TFL1-likegene, reduces the juvenile phase inapple. J Am Soc Horticult Sci 131:74–81

Lehti-Shiu MD, Adamczyk BJ, Fernandez DE (2005) Expression ofMADS-box genes during the embryonic phase in Arabidopsis.Plant Mol Biol 58:89–107

Litz RE, Raharjo SHT (2005) Sapindaceae. In: Litz RE (ed) Biotechnologyof fruit and nut crops. CABI, Wallingford, pp 627–636

Magdalita PM, Laurena AC, Yabut-Perez BM, Mendoza EMT, VillegasVN, Botella JR (2002) Progress in the development of transgenicpapaya: transformation of Solo papaya using ACC synthase anti-sense construct. Acta Horticult 575:171–176

Mathews H, Litz RE, Wilde DH, Merkel S, Wetzstein HY (1992) Stableintegration and expression of β-glucuronidase and NPT II genes inmango somatic embryos. In Vitro Cell Dev Biol Plant 28:172–178

Mathews H, Litz RE, Wilde HD, Wetzstein HY (1993) Genetic trans-formation of mango. Acta Horticult 341:93–97

Matsumoto TK (2006) Genes uniquely expressed in vegetative andpotassium chlorate induced floral buds of Dimocarpus longan.Plant Sci 170:500–516

Matzke AJM, Matzke MA (1998) Position effects and epigenetic si-lencing of plant transgenes. Curr Opin Plant Biol 1:142–148

Meins F (2000) RNA degradation and models for post-transcriptionalgene silencing. Plant Mol Biol 43:261–273

Menzel CM (2002) The lychee crop in Asia and the Pacific. FAO,Regional Office for Asia and the Pacific. Bangkok, Thailand

Menzel CM, Huang X, Liu C (2005) Cultivars and plant improvement.In: Menzel CM, Waite GK (eds) Litchi and longan: botany, pro-duction and uses. CABI, Wallingford, pp 59–86

Meyerowitz EM, Smyth DR, Bowman JL (1989) Abnormal flowers andpattern formation in floral development. Development 106:209–217

Murashige T, Skoog F (1962) A revised medium for rapid growth andbioassays with tobacco tissue cultures. Physiol Plant 15:473–497

Nolan T, Hands RE, Bustin SA (2006) Quantification of mRNA usingreal-time RT-PCR. Nat Protoc 1:1559–1582

Omar AA, Dekkers MGH, Graham JH, Grosser JW (2008) Estimationof transgene copy number in transformed citrus plants by quanti-tative multiplex real-time PCR. Biotechnol Prog 24:1241–1248

Peirson SN, Butler JN, Foster RG (2003) Experimental validation ofnovel and conventional approaches to quantitative real-time PCRdata analysis. Nucleic Acids Res 31:e73

Pfaffl MW (2001) A new mathematical model for relative quantifica-tion in real-time RT-PCR. Nucleic Acids Res 29:e45

Pfaffl MW, Horgan W, Dempfle L (2002) Relative expression software tool(REST) for group-wise comparison and statistical analysis of relativeexpression results in real-time PCR. Nucleic Acids Res 30:e36

Raharjo SHT, Litz RE (2007) Somatic embryogenesis and plant regen-eration of litchi (Litchi chinensis Sonn.) from leaves of maturephase trees. Plant Cell Tissue Organ Cult 89:113–119

Raharjo SHT, Witjaksono, Gomez-Lim MA, Padilla G, Litz RE (2008)Recovery of avocado (Persea americana Mill.) plants transformedwith the antifungal plant defensin gene pdf1.2. In Vitro Cell DevBiol Plant 44:254–262

Ribas AF, Kobayashi AK, Pereira LFP, Vieira LGE (2006) Production ofherbicide-resistant coffee plants (Coffea canephora P.) viaAgrobacterium tumefaciens-mediated transformation. Braz Arch BiolTechnol 49:11–19

Riechmann JL, Meyerowitz EM (1997) MADS domain proteins inplant development. Biol Chem 378:1079–1101

Samach A, Kohalmi SE, Motte P, Datla R, Haughn GW (1997)Divergence of function and regulation of class B floral organidentity genes. Plant Cell 9:559–570

Schwarz-Sommer Z, Hue I, Huijser P, Flor PJ, Hansen R, Tetens F,Lönnig W, Saedler H, Sommer H (1992) Characterization of theAntirrhinum floral homeotic MADS-box gene deficiens: evidencefor DNA binding and autoregulation of its persistent expressionthroughout flower development. EMBO J 11:251–263

Shepherd CT, Moran Lauter AN, Scott MP (2009) Determination oftransgene copy number by real-time quantitative PCR. MethodsMol Biol 526:129–134

StamM, de Bruijn R, Kenter S, van der Hoorn RAL, van Blokland R, MolJNM, Kooter JM (1997) Post-transcriptional silencing of chalconesynthase in Petunia by inverted transgene repeats. Plant J 12:63–82

Subhadrabandhu S, Stern RA (2005) Taxonomy, botany and plantdevelopment. In: Menzel CM, Waite GK (eds) Litchi and longan:botany, production and uses. CABI, Wallingford, pp 25–34

Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De PaepeA, Speleman F (2002) Accurate normalization of real-time quan-titative RT-PCR data by geometric averaging of multiple internalcontrol genes. Genome Biol 3:R0034

Viruel MA, Hormaza JI (2004) Development, characterization andvariability analysis of microsatellites in lychee (Litchi chinensisSonn., Sapindaceae). Theor Appl Genet 108:896–902

Weigel D, Meyerowitz EM (1994) The ABCs of floral homeotic genes.Cell 78:203–209

WongWS, Li GG, NingW, Xu ZF, HsaioWLW, Zhang LY, Li N (2001)Repression of chilling induced ACC accumulation in transgeniccitrus by overproduction of antisense 1-aminocyclopropane-1-car-boxylate synthase RNA. Plant Sci 161:969–977

Wu Y, Yi G, Zhou B, Zeng J, Huang Y (2007) The advancement ofresearch on litchi and longan germplasm resources in China. SciHortic 114:143–150

Yao J-L, Dong YH, Morris B (2001) Parthenocarpic apple fruit produc-tion conferred by transposon insertion mutations in a MADS-boxtranscription factor. Proc Natl Acad Sci U S A 98:1306–1311

Ying ZT, Yu X, Davis MJ (1999) New method for obtaining transgenicpapaya plants by Agrobacterium-mediated transformation of so-matic embryos. Proc Fla State Hortic Soc 112:201–205

Zhong H-Y, Chen J-W, Li C-Q, Chen L, Wu J-Y, Chen J-Y, LuW-J, Li J-G (2011) Selection of reliable reference genes for expression studiesby reverse transcription quantitative real-time PCR in litchi underdifferent experimental conditions. Plant Cell Rep 30:641–653

PADILLA ETAL.